Sanity check anyone? (was Re: [squeak-dev] The Inbox: System-cbr.375.mcz)

Nicolas Cellier nicolas.cellier.aka.nice at gmail.com
Mon Sep 13 07:31:36 UTC 2010


I recently had several crash dump too with same configuration.
I had a few changes related to Xtream though...

Nicolas

2010/9/13 Frank Shearar <frank.shearar at angband.za.org>:
> Clean 4.2 image, running on Squeak4.0.1.exe. Update to the latest Trunk
> (#10518), load System-cbr.375 and... it all just works.
>
> The failing image has been incrementally updated starting sometime soon
> after we got the Squeak 4.0 installer up and running, but it otherwise
> hasn't had any local changes.
>
> frank
>
> On 2010/09/13 01:59, Casey Ransberger wrote:
>>
>> I can't seem to reproduce this issue (which of course is a far cry from
>> saying that it isn't real.) Would any kind Squeaker (preferably running
>> the same Windows VM that Frank is using?) try loading System-cbr.375.mcz
>> into an updated trunk image to give us a sanity check?
>>
>> I'm not accusing the VM of any malfeasance (because I probably just
>> messed up somewhere) but it would be nice to have someone try it with a
>> similar config to what Frank has just to be sure.
>>
>> On Sun, Sep 12, 2010 at 1:49 PM, Casey Ransberger
>> <casey.obrien.r at gmail.com <mailto:casey.obrien.r at gmail.com>> wrote:
>>
>>    Weird. I have no idea why that would happen; all that code does is
>>    patch the method that sets the default fonts and run it. I used the
>>    same font that the system uses currently for code (but in a couple
>>    sizes.) It worked okay on my box, so something must be different.
>>
>>    What version of the image did you load that into?
>>
>>    I will try to find some time to look at the stack trace you
>>    supplied; thanks for going to the effort to do that, and I'm very
>>    sorry that my changes caused trouble for you.
>>
>>    On Sep 12, 2010, at 5:25 AM, Frank Shearar
>>    <frank.shearar at angband.za.org <mailto:frank.shearar at angband.za.org>>
>>    wrote:
>>
>>     > On 2010/09/11 21:13, Casey Ransberger wrote:
>>     >> Okay, Bert's suggestion that I add a comment to get the
>>    postscript to
>>     >> run worked. Pushing this one into trunk will give us nice  modern
>>     >> looking fonts in buttons, flap tabs, and balloon help.
>>     >>
>>     >> Feedback would be appreciated. I went for the nearest
>>    approximation of
>>     >> the old font sizes, but some folks might find these to be a bit
>>    small
>>     >> (they work fine for my eyes, but YMMV.) We could always bring
>>    the sizes
>>     >> up a notch.
>>     >>
>>     >> The previous package, System-cbr.374, can be trashed. It had the
>>    same
>>     >> changes, but as I didn't change the postscript, it was not run
>>    on load.
>>     >
>>     > I'm running on a clean, just-updated trunk, version #10508, on a
>>    Squeak4.0.1 from Mar 28 2010, Windows XP.
>>     >
>>     > I open up a Monticello browser, and load the mcz. I'm told
>>    "Initializing System-cbr.375" (the progress bar itself is solid
>>    white - no progress, in other words). I watch memory increase
>>    steadily. Eventually I hit a low space warning at ~500M usage. After
>>    that, the VM crashes:
>>     >
>>     > "Sorry but the VM has crashed.
>>     > Reason: out of memory
>>     >
>>     > Current byte code: 20
>>     > Primitive index: 0"
>>     >
>>     > The attached crash.dmp shows my original attempt as well as a
>>    subsequent attempt, just to make sure I could duplicate the problem.
>>     >
>>     > frank
>>     > <crash.dmp>
>>     >
>>
>>
>>
>>
>> --
>> Casey Ransberger
>>
>>
>>
>>
>
>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: crash.dmp
Type: application/octet-stream
Size: 14663 bytes
Desc: not available
Url : http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20100913/9324cdb1/crash.obj


More information about the Squeak-dev mailing list